Residual posttraumatic stress disorder symptoms after provision of brief behavioral intervention in low‐ and middle‐income countries: An individual‐patient data meta‐analysis

Abstract Background To address shortages of mental health specialists in low‐ and middle‐income countries, task‐shifting approaches have been employed to train nonspecialists to deliver evidence‐based scalable psychosocial interventions. Problem Management Plus (PM+) is a brief transdiagnostic nontrauma focused intervention for people affected by adversity. This study reports on the capacity of PM+ to address specific symptoms of posttraumatic stress disorder (PTSD). Methods Individual patient data from three randomised controlled trials were combined and analysed to observe the impacts of PM+ (n = 738) or enhanced treatment as usual (ETAU) (n = 742) interventions on specific PTSD symptoms at posttreatment and 3‐month follow‐up. The PTSD‐Checklist for DSM‐5 (PCL‐5) was used to index PTSD symptoms, and presence of each symptom was defined as moderate severity (score ≥ 2 on individual items). Results The average PCL‐5 score at baseline was 26.1 (SD: 16.8) with 463 (31.3%) scoring above 33, indicative of a diagnosis of PTSD. Following intervention, 12.5% and 5.8% of participants retained a score greater than 33 at postassessment and follow‐up, respectively. There was greater symptom reduction for PM+ than for ETAU for most symptoms. Hyperarousal symptoms were the most common residual symptoms after PM+, with more than 30% of participants reporting persistent sleep disturbance, concentration difficulties, and anger. Conclusion PM+ led to greater reduction in symptoms relating to re‐experiencing and avoidance. The evidence indicates that strategies focusing on hyperarousal symptoms including sleep, concentration, and anger difficulties, could be strengthened in this brief intervention.


| INTRODUCTION
People affected by war, conflict, and humanitarian crises disproportionately live in low-and middle-income countries (LMICs), and they are at increased risk for psychological problems (Barbui & Tansella, 2013;Charlson et al., 2019). This is not surprising because these vulnerable populations can be exposed to significant adversity, including poverty, sexual violence, torture, war, and displacement.
Despite the prevalence of mental health problems in LMICs, including people living in humanitarian contexts, it is estimated that at least 90% of adults with mental health needs do not receive minimum adequate care (Chisholm et al., 2016). Programs exist that can address mental health conditions in these typically poor resource settings (Morina et al., 2017a(Morina et al., , 2017b, but they are often not implemented because (a) these interventions tend to only target a single diagnostic outcome (e.g., posttraumatic stress disorder (PTSD; (Neuner et al., 2008)), (b) are generally resource and time intensive (Bolton et al., 2007), and (c) often require mental health specialists who are lacking in LMICs (Bass et al., 2013). These factors preclude many LMICs and humanitarian responses from implementing muchneeded mental health services that their populations require (Patel et al., 2018;Tol et al., 2012). This situation has led to increased recognition that addressing the mental health and psychosocial needs of people in humanitarian crisis settings requires transferring a portion of mental health delivery to trained and supervised nonspecialist workers. This shift is appealing in low resource settings because it overcomes the limitation of the scarcity of mental health specialists, reduces costs, and can minimize stigma associated with specialist mental health care (Patel et al., 2018). Increasingly, programs have utilized these "taskshifting" approaches in which local providers who lack formal mental health qualifications are trained to deliver structured manualized interventions for common mental disorders. One meta-analysis found these interventions are moderately effective in LMICs (Singla et al., 2017).
The World Health Organization (WHO) has recently adopted this approach by developing a range of mental health interventions that are intended to address common mental disorders (i.e., anxiety and depression) in populations affected by adversity. Inherent in this approach is that these interventions are brief, affordable, can be easily trained to nonspecialists, and in this manner can be scaled up to achieve maximum reach in an LMIC and with limited resources.
The first program of this type developed by the WHO was titled Problem Management Plus (PM+), a transdiagnostic intervention that involves five sessions that teach people skills that have been shown to be the most effective in reducing common mental disorders in controlled trials . Specifically, PM+ teaches nonspecialist providers to train people with common mental disorders skills in arousal reduction, problem-solving, behavioral activation, and strengthening social supports. PM+ delivered to individuals has been shown to be effective in reducing psychological distress in adverse settings. The two large trials comprised women affected by gender-based violence in Kenya  and men and women in in a conflict-affected region in Pakistan . Additionally, PM+ has been delivered in a group format in Pakistan, which has also been shown to be effective relative to treatment as usual (Rahman et al., 2019).
Despite the overall efficacy of PM+ in reducing psychological distress and symptoms of common mental disorders, there is a need to better understand how it impacts specific psychological symptoms following adversity, such as the experiences of people who have lived through interpersonal trauma or humanitarian crises. One treatment study of female veterans with PTSD that employed trauma-focused cognitive behaviour therapy (TF-CBT) found that hyperarousal symptoms, and especially sleep problems and irritability, were most resistant to change with treatment (Schnurr & Lunney, 2019). This finding accords with a trial of female rape survivors with PTSD, and found that sleep problems were most resistant to TF-CBT (Larsen et al., 2019). The finding that many people continue to experience functional difficulties following remission of PTSD diagnosis  points to the need to better understand psychological difficulties that persist after treatment. Although some inroads have been made in relation to trauma-focused psychotherapies, there is a need to understand how scalable interventions, such as PM+, address specific PTSD symptoms in humanitarian and development contexts where the prevalence of PTSD is elevated (Charlson et al., 2019). This is relevant for PM+ because although it is not a trauma-focused intervention, previous trials have shown that PM+ leads to significant reductions in PTSD severity, in adversity-affected LMICs and conflict settings (Rahman et al., , 2019. In both situations, many people in these category do suffer from PTSD, or experience events that could lead to PTSD. In turn, understanding the residual symptoms of PTSD following PM+ would provide information on how the intervention reduces PTSD symptoms as well as inform the development of future scalable interventions. To this end, this study conducts an individual-patient data meta-analysis (IPD-MA) of the three large controlled trials that have been published to determine which symptoms of PTSD are most persistent after provision of PM+ Rahman et al., , 2019. It further aims to inform the potential of PM+ to be a useful brief intervention across humanitarian and adversity settings where PTSD is considered one of the common mental disorders experienced. All participants provided written informed consent before taking part in the trials.

| Context and participants
Between November 2014 and August 2016 three RCTs were conducted to evaluate the effectiveness of PM+ in reducing the levels of psychological distress experienced by participants. A secondary outcome of each trial included PTSD symptoms. Data from these studies were combined into a single data set.
The study characteristics are presented in Table 1. In the Pakistan individual PM+ study, participants (N = 346) were recruited from three primary health care centres in the conflict-affected urban city of Peshawar . Participants received clinical assessments from their physicians; participants who were deemed to be suffering from psychological distress were referred to the study.
The Pakistan group-based PM+ study was conducted in two rural council districts of Swat (Rahman et al., 2019), and adult women (N = 612) were randomly screened in the community and invited into the study if they screened positively for psychological distress. The Kenya study recruited female participants (N = 522) through random sampling in peri-urban neighbourhoods of Nairobi, of whom 81% had experienced gender-based violence . In all three studies the initial screening procedure used the General Health Questionnaire-12 (GHQ-12; (Goldberg & Williams, 1988;Minhas & Mubbashar, 1996)) to identify psychological distress and the WHO-Disability Assessment Schedule 2.0 (WHODAS 2.0), which is a 12-item measure of general functioning (WHO, 2010). The common inclusion criteria across all three studies were (i) adults aged 18-60 years, (ii) score of ≥3 on the GHQ-12, and (iii) score of ≥17 on the WHODAS 2.0. The exclusion criteria were identical across all three studies: (i) imminent risk for suicide, (ii) severe mental illness (e.g., psychotic disorders, substance use disorders), and (iii) significant cognitive and neurological impairment (e.g., severe intellectual disability).

| Measure
PTSD symptoms were measured using the PTSD Checklist for DSM-5 (PCL-5) (Weathers et al., 2013). The PCL-5 is a 20-item checklist corresponding to the 20 symptoms for PTSD as defined in DSM-5.
There are five items corresponding to intrusion symptoms, two to avoidance symptoms, seven to negative alterations in cognitions and moods, and six to alterations in arousal and reactivity. Items are rated on a 5-point Likert scale (0 = not at all, 4 = extremely) and total scores are obtained by summing the responses for individual items (range: 0-80). Across all three studies, the PCL-5 was adapted to ask for symptoms in the past week (instead of month) to enhance sensitivity to change. The PCL-5 measure was adapted in all three studies in accordance with gold-standard translation practices, with each item being translated and back-translated by accredited translators (Bontempo, 1993). In Kenya, items were translated into Kiswahili and in Pakistan, items were translated into Urdu. The comprehensibility of the adapted versions were piloted in feasibility trials which took place before the RCTs .
The psychometric properties of the PCL-5 have been widely studied and has been shown to have high levels of diagnostic accuracy and internal consistency (Blevins et al., 2015;Bovin et al., 2016), including in LMIC (Mughal et al., 2020). Symptoms were rated as present if the respondent indicated that they were at least moderately bothered by them, as indicated by a score ≥2 on individual items (Weathers et al., 2013). Symptom retention was defined as the conditional probability of retaining a symptom at posttreatment and 3-month follow-up compared to before receiving treatment. A probable diagnosis of PTSD is made if the total score on the PCL-5 was ≥33 (Weathers et al., 2013).

| Statistical analysis
Means and SD are reported for continuous variables while frequencies and percentages are reported for categorical variables.
Missing outcome data at postintervention and follow-up assessments were estimated using multiple imputation (MI); data were assumed to be missing-at-random. As the PCL-5 results are scored on a Likert scale, MI was conducted by chained equations using ordered logistic regression (mi impute chained ologit in Stata). We analysed the effects of PM+ versus ETAU on the conditional probability that participants retained symptoms following intervention using a multilevel logistic regression, with random effects for each trial; separate models were run for each of the 20 PTSD symptoms. A completecase sensitivity analysis was conducted to explore robustness of results from the MI using only participants who completed the post and follow-up assessments. Results are reported as the estimated odds ratio (OR) and the 95% confidence intervals (CI) for each symptom are presented. All analyses were conducted using Stata 13 (StataCorp., 2013).

| RESULTS
Descriptive information about study participants is presented in   Tables 3 and 4 present the proportions of individuals who retained symptoms following intervention (Table 3) and at 3 months (

| DISCUSSION
We examined the residual symptoms of PTSD following a lowintensity intervention across urban, peri-urban, and rural samples in three major trials in two LMICs, Kenya and Pakistan. We found that PM+ resulted in higher alleviation of symptoms when compared to ETAU. This was most notable across avoidance, cognition/mood, and arousal symptoms. These trends were similar at both postintervention assessments and 3-month assessments.
PM+ had a significant impact on the alleviation of avoidance symptoms when compared to ETAU immediately following the program. Moreover, there was approximately a 75% reduction of re-experiencing symptoms (e.g., intrusive memories, nightmares, flashbacks). These are interesting findings insofar as PM+ is not a trauma-focused intervention, and does not directly target emotional processing of trauma memories or avoidance of trauma reminders.
Trauma-focused therapies purportedly lead to reduction of PTSD symptoms because they promote emotional processing of trauma memories and associated emotions through explicit exposure to trauma memories, which can facilitate extinction learning of previously learnt association of fear (Foa, 2006). It is possible that the use of problem-solving, behavioural activation, and social support in PM+ prompts participants to engage in activities that facilitate emotional processing because of discussion about core problems or approaching previously avoided situations that are reminiscent of the traumatic experience. It is also possible that the focus on problemsolving skills in PM+ may address aspects of avoidant behaviors; for example, problem-solving approaches to being reluctant to shop at a local market because it reminds a person of being at a market when it was bombed may lead to proactive strategies to return to markets, thereby implicitly engaging in in vivo exposure activities. Further,

TF-CBT is intended to involve integration of new information that
corrects excessively negative appraisals about the person and their world (Ehlers & Clark, 2000). Problem-solving and behavioural activation strategies may result in beliefs about danger of one's environment or one's inadequacy may be alleviated by these strategies, which could also reduce avoidance behaviors. The finding that the superior effect regarding avoidance thoughts was not observed at 3 months may point to the potential benefit of providing booster sessions to maintain the relative benefits of PM+.
The finding that negative cognitive and mood symptoms were better addressed by PM+ than ETAU may also be attributed to the behavioral activation components of PM+ intervention. There is considerable evidence that behavioral activation reduces low mood (Ekers et al., 2014), arguably because it prompts individuals to engage in potentially pleasurable activities that can increase their sense of reward. Behavioral activation may additionally promote self-efficacy through successful completion of tasks, which in turn can improve self-esteem and self-related appraisals (Benight & Bandura, 2004).
Further, the act of problem-solving itself can promote self-mastery because it facilitates the perception that one can influence outcomes, thereby potentially reducing learned helplessness and enhancing mood (Cassidy & Long, 1996).
At follow-up there were fewer PM+ participants with arousal symptoms, relative to those who received ETAU, including hypervigilance, startle response, concentration difficulties, anger, and sleep disturbance. One of the core strategies of PM+ aims to reduce arousal via simple techniques, primarily deep breathing strategies.
Previous research has found a number of breathing-based techniques helpful for persons with PTSD, specifically in reducing hyperarousal symptoms (Seppälä et al., 2014). Moreover, meta-analyses indicate that slow breathing is efficacious in reducing a range of arousal symptoms, including anger, sleep disturbance, and poor concentration (Zaccaro et al., 2018). Thus it is unsurprising that teaching participants diaphragmatic breathing during the course of PM+ led to reductions in arousal-related problems of PTSD.
Although PM+ resulted in reductions of arousal-related problems, hyperarousal symptoms tended to be the most resistant to remission across both arms. Consistent with prior trials that have employed trauma-focused psychotherapy for PTSD (Larsen et al., 2019;Schnurr & Lunney, 2019), this study found that sleep problems was the most retained symptom, followed by hypervigilance, startle response, anger, and concentration difficulties 3 months after PM+. This finding across treatment trials underscores the recognition that sleep disturbance is a common transdiagnostic problem for many people with psychological distress which might assume independent status as distinct from other presenting diagnoses (Spoormaker & Montgomery, 2008;Zayfert & DeViva, 2004).
Sleep impairment is a commonly persistent problem after treatment for depression (Menza et al., 2003), suggesting that lingering sleep difficulties in participants may be associated with ongoing depression. It is also worth noting that many participants in the current trials The reduction of symptoms for participants who received ETAU is additionally noteworthy. Several explanations may be considered to account for this pattern. First, ETAU involved providing participants with psychoeducation following their baseline assessment. On its own, this may have been beneficial, as supported by evidence that psychoeducation can reduce some PTSD symptoms (Hadar Lubin et al., 1998;Swan et al., 2017). Second, there is evidence that repeated assessments themselves may result in reduction of symptom reporting, which formed part of the methologies in these trials (Tarrier et al., 1999).  (Hamdani et al., 2021) and also in high-income countries with advanced health resources (de Graaff et al., 2020).

| LIMITATIONS AND FUTURE RESEARCH
We note several methodological limitations. First, two of the three included studies only enrolled women, which may limit the generalizability of these findings beyond female populations. Recruitment of males can be difficult in humanitarian crisis settings (Affleck et al., 2018), however there are sex differences in both PTSD and in treatment response (Olff, 2017). Therefore, continued understanding about how gender impacts symptom remission remains important.
Second, the trial participants were selected due to heightened levels of psychological distress and no formal diagnostic assessments via structured clinical interviews were conducted. Third, there was a large difference on average baseline scores as well as the number of participants who scored ≥33 on the PCL-5 in Kenya/Peshawar and Swat. This may be attributed to the different contexts and participants selected; Kenya data was based only on women who had a reported history of gender-based violence, while the Peshawar study recruited participants in a high-stress environment where acute conflict was commonplace. In contrast, the Swat trial participants were recruited from rural districts affected by chronic conflict and natural disaster. In the context, we also note that there is no convergent evidence regarding the recommended cut-off scores on the PCL-5 to identify probable PTSD in humanitarian settings. Fourth, trials were not able to collect detailed information about usual care treatments that participants received instead of PM+. Being able to compare the ETAU interventions across studies to look at commonalities and differences would allow for a better understanding of the patterns and results observed across the three trials. Lastly, we recognize that meeting diagnostic criteria was not a prerequisite for entry into the studies.

| CONCLUSION
As transdiagnostic programs are increasingly delivered by lay providers, it is important to understand how these programs impact on specific symptoms across a range of disorders. This is the first study to investigate the direct effects of PM+ on the residual symptoms on PTSD. Although PM+ was designed as a transdiagnostic intervention and is not specifically a trauma-focused treatment, it is a valuable finding that it was still effective to reduce many core PTSD symptoms, including re-experiencing and avoidance symptoms. Importantly, PM+ was less successful in reducing certain hyperarousal symptoms, which may reflect ongoing problems associated with living in contexts of adversity, poverty, and humanitarian crisis. Identifying residual symptoms that persist after PM+ will assist future development of programs that could address lingering problems that people experience after receiving brief transdiagnostic programs. AKHTAR ET AL.